Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: SSDF Rating List 2006-02-10

Author: Sandro Necchi

Date: 10:42:51 02/11/06

Go up one level in this thread


On February 11, 2006 at 08:59:51, Albert Silver wrote:

>On February 11, 2006 at 04:16:28, Sandro Necchi wrote:
>
>>On February 11, 2006 at 03:59:05, Uri Blass wrote:
>>
>>>On February 11, 2006 at 02:01:44, Peter Stojanov wrote:
>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>Why they test the engines that have default opening books with a general
>>>>book,sorry but I don't take their results and rating list seriously.
>>>
>>>Thee ssdf use default opening book.
>>>
>>>I do not see a reason not to take results seriously because the testers do not
>>>use the default opening book.
>>>
>>>There are people who are interested in results with default opening books and
>>>there are people who are interested in results of the engines not with the
>>>books.
>>
>>true, but if you change the book it is a different player, so the name of the
>>program should be changed to show this.
>>
>>I know some people would not agree (who cares), but it is like to force a human
>>player to play different openings...it is not the same player anymore as results
>>are concerned.
>
>>Sandro
>
>I agree. When I test versions and parameters, I am testing the engine, thus I
>use an opening suite to annul the effect the openings might have on the result.
>
>However, if my intention were to test the better program, I'd have to use the
>book included, since presumably that is what will bring the greatest results. I
>don't think one should eliminate the investment made to have a custom-made
>top-notch book such as yours. A strong well-tuned opening repertoire is a
>significant part of any master's or grandmaster's playing performance.
>
>Possibly worse, is to use a book that is fine-tuned for an engine with a very
>specific style (such as Junior 9), and impose it on an engine with a completely
>different playing style. The book may give great results for the engine it was
>designed for, but can easily give much worse results for an engine that is not
>addapted for it.

I agree as every program has it's playing style and therefore preferred
lines/openings.
To force someone to play what it does not like is always bad from results point
of view; especially for the programs which still have several weaknesses...

>                                      Albert

Sandro



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.